Aug 3, 2026 earnings call — Quarter results, credit stress, software AI review
Adjusted net income per share was $0.22, up from an $0.18 per share loss last quarter.
Revenue $0.2B (est $0.2B) · EPS $0.34 (est $0.33)
Key quotes
"Well, GBDC's net realized and unrealized losses for the quarter, the $0.12 per share of loss, they arose primarily from a small number of junior debt and equity positions and not from the core debt portfolio."
"The results of our internal AI risk analysis showed that less than 10% of our software portfolio was subject to elevated AI disruption risk. The third party consultants AI risk assessment concluded that fewer than 3% were at elevated risk."
"The company bought back 1.1 million shares at a weighted average price of $12.90 per share or an approximate 10% discount to our March 31, 2026, net asset value."
Analyst Q&A 5
Capital allocation and home services weakness
Q (Finian O'Shea · Wells Fargo): Analyst asked how much the focus on buybacks and deleveraging was driven by the quality and price of new investments, and whether that preference will continue as leverage falls. He also asked about recent credit pressure in home services and whether it reflects broader themes like healthcare roll-ups.
A: They weigh trade-offs between repurchases, new investments, and staying in target leverage. They slowed new investing last quarter because payoffs were slow, to support repurchases and some deleveraging, and expect payoffs to increase giving more flexibility. For home services, slower moving activity since 2022 reduced demand, so pressure may continue for a while but should self-cure over time.
"I think the payoffs are going to increase, and that's going to give us a lot more flexibility to be able to play more in new investing activity and simultaneously continue to achieve our leverage goals and our repurchase goals." — David B. Golub (CEO)
Deal terms and software AI risks
Q (Kenneth Lee · RBC): Analyst asked for more detail on deal activity and the pickup in spreads and terms versus the March quarter. He also asked about the sequential rise in lower risk grades in the software loan portfolio and common themes.
A: M&A for private equity remained constrained in Q1 and only recovered a bit in Q2, so competition limited spread widening. They saw a 25 basis point to 50 basis point improvement in pricing and some better terms and leverage, and expect more improvement as M&A recovers. On software, AI is changing coding costs and creating winners and losers; they found a small subset of software loans vulnerable to AI and are reflecting that in valuations.
"I mentioned in the prepared remarks about a 25 basis point to 50 basis point improvement in pricing, and there's also been some improvements in terms and in leverage levels." — David B. Golub (CEO)
Identifying portfolio tail and growth quality
Q (Paul Johnson · KBW): Analyst asked whether the troubled 'tail' in the portfolio is mostly identified or still growing, and where the industry stands in the normalization. He also asked about the quality of growth in sponsor-backed portfolios, and how much is organic versus M&A-driven.
A: They view this as a credit cycle with elevated stress that creates winners and losers, and they emphasize early identification of problems. They believe they are mostly through identifying the credits in their portfolio that will have challenges, while some other managers may lag. Revenue and EBITDA growth in portfolio companies is moderate—resilient but slower than post-COVID—and pent-up M&A should pick up if uncertainty around rates and energy eases.
"I think we're in at a phase where we're most of the way towards having identified the credits in our portfolio that are going to have challenges." — David B. Golub (CEO)
Software nonaccruals, AI risk, dividend policy
Q (Robert Dodd · Raymond James): Analyst asked whether the three new software nonaccruals were all from companies with elevated AI risk or if other themes existed, and why the third-party AI-risk assessment differed from the company's internal view. He also asked about the likelihood of a dividend policy change given spreads and base rates.
A: There are multiple reasons for software underperformance, not just AI; some cases involve acquisitions and integration issues, while AI will remain a meaningful factor. He said the internal and third-party analyses are more similar than they appear and both show low numbers, and that Golub's portfolio is less vulnerable because of long-standing underwriting choices favoring enterprise systems. On dividends, they aim to pay an amount that serves shareholders while holding steady NAV and not changing the dividend often; recent trends were helpful but they will keep watching.
"The reality, if you ask me, is that if we looked across the software industry, the proportion of software companies that are going to be vulnerable to AI-related elevated risks, it's a lot higher than what's in our portfolio." — David B. Golub (CEO)
Secondary market buys and Q2 NAV moves
Q (Ethan Kaye · Lucid): Analyst asked how large the secondary market purchases were this quarter, whether more opportunities remain, and at what discounts those purchases were done. He also asked why GBDC saw less spread-related NAV pressure in Q2 than a peer.
A: They have a long-standing sales and trading desk and use it to buy loans when attractive; calendar year-to-date Golub Capital sales and trading exceeded $2 billion and produced some opportunities, but for GBDC it was not a meaningful source of new investments in Q2. They saw little spread-related valuation movement in Q2 (that hit mostly in Q1), and the main story was stabilization with credit-related valuation changes across the industry in Q2.
"In the calendar year-to-date, the Golub Capital sales and trading activity has exceeded $2 billion." — David B. Golub (CEO)